AltME groups: search

I think the way to go is to convert the source into a DLL and recompile. 
I would like to try a bit later - but first - bread.

I tried it recently and made a DLL which I interfaced with rebol 
very quickly.

err... I think not all... because I was missing a file winmm.lib 
(which I found in masm32 distribution), and mmreg.h (which I found 
on the web, though don't know if if was right version)

Winmm.dll is the MCI interface in Windows. With that interface you 
can control many multimedia stuff (wave form playing, video playing, 
recording...)

It has two interface : C++ level and string command. The last one 
sounds easy to interface with view/pro. But, eh, I never played with 
DLL interfacing in rebol.
So I let that to "make routine!" guru ;-)

I can imagine following kind of UI being done with View ... except 
video and audio playback for now ...  http://www.mythtv.org/modules.php?name=MythFeatures

I'm working on a sound library, for /Core and /View, with - for now 
- AIFF / 8SVX / AU / WAVE (PCM/float/u-law/A-law) support (read and 
write). Any suggestion for another usefull format to support?

REBOL cuts out the bureacracy of UML and implements direct.  Schler-Mellor 
and Grady Brooch have some "noble" goals (but the devil's in the 
implementation), but have never faced anything as powerful as REBOL 
yet for implementation.  I saw Mellor talk once, and he did a great 
talk on "orthogonal behavior" which is apropos.  I think the way 
to utilize their mess is to do it simpler in REBOL, call it UMLR 
and teach them the future.

I will soon now post some doc(s) for brainstorming - well, maybe 
not docs, just a brainstorming itself ... I need to find myself and 
categorise my thoughts and need help on that ... of course I am studying 
various workflow, business process papers etc ...

... but maybe just because UML is here for real object oriented languages 
like Java, and once we use Rebol, we don't need such beurocracy to 
express relations ...

you need to define the lexicon of workflow.  keep it simple, and 
follow the flow

(Vincent...I have the CD's from TIA EIA... with sample C code...and 
some other code as well...;-)

if you are in Europe, then the GSM standards might be most useful 
to you, and I would have to look them up.

sound:// need streaming capabilities, for both music and VoIP

Just reading around and getting up-to-date... I found RAMBO. I send 
a bug report with the "Feedback" thing on RT's homepage. Should I 
post it again with RAMBO? Maybe I'll get better response?

Geomol: i'd suggest using RAMBO and not email feedback for bugs. 
this frees up some time for Carl.

About 50k of REBOL/Core source. Build as a relative level and an 
index level on top of that - top-most a full relational DB. I've 
also started a SQL interface, but I'm not sure, if I wanna finish 
that part. :-)

related to what? Rebol? Well, there were those times, in 1996, when 
Carl tried to resurrect amiga at Viscorp ... then he left, I read 
his paper about "dream system" - first scripting, then OS, and I 
believed into it. Well, we are a bit late with the original plan, 
but I still try to believe :-)

He was right. And many of those words still stands. Only today the 
complexity is even worse, the need for CPU Power, RAM and HD space 
is even higher.

on UML.


my take on that is it can consume a lot of time and energy.... as 
engineers try to 'perfect' their model.


Personally, I use FLiP process ( http://www.fusebox.org/index.cfm?&fuseaction=methodology.steps
) or Fusebox for short.

It makes it easier for end users to understand.

I've done the exact same thing and requested help here too, so I 
feel ya. ;-)

:-) I start debugging in a typical way for me ... started to commented 
out the code ... but then I reached the level, where the code would 
not obviously work, and so I went to file-manager to check things 
out ...

I just want to say that I tested out Doc's NTLM library and it worked 
very well today.  I was successful in testing this out on our corporate 
network.  If Altme had something like this we could use ALTME from 
behind the corporate firewall.

yesterady, I was rather pleasantly surprised. At XidysComp, we use 
http://www.MoneyS3.CZaccounting system, but mainly for our PC sales 
store, invoicing, connection to POS system etc. It has some limitation, 
as e.g. one product, e.g. Seagate 80GB, 7200 can have just one store 
number and one reference number to dealer. Well, but we buy the same 
product from more than one dealer.

So we asked money.cz representative to come and explain few things, 
and we've got Director himself. He agreed that there is some limitation 
and was surprised, when I showed him Rebol, and small parser for 
few dealer pricelist. He said that was exactly his idea someone would 
do something like that and that it could be sold. He liked the way 
we think (the rebol way) and he wants to teach us more about S3 to 
become sales partner for some customers ...

It is really nice when oportunities rise that way and that someone 
can see potential in small, flexible tools .....

and BTW, maye IN-INTERVAL? can be useful to anyone else too? checks 
if a value is inside an interval, assuming a ring. can be easily 
used for rings mod 2^n where n is multiple of 8, just using binaries. 
see http://www.colellachiara.com/soft/Libs/chord.r

and so on , and so on  ... :^)

yup, old, and still on the list for the next rev.

How far are we from AltME for OS X? I'm only on this PC for about 
an hour a day, and I'd love to spend some of the other time on AltME.

I have some clients who would benefit from Altme, if it was possible 
to control which users can be seen by others. ie not just private 
groups, but private users. This is handy for organising sales channel 
partners, where you want them to feel they have a direct line to 
the people in your office, (but can't see the other sales partners), 
and for staff they can maintain & respond to all channel partners 
requests from one simple tool

Wow, did someone forget to change the monkey diapers in the peanut 
gallery?
You and Graham sure seem ornery.

Reichart, ever since Altme came out, you've been speaking of the 
"list".. not sure how many years so far.  What was the last update, 
scrolling?  If you don't want to update Altme, that's your perogative, 
but the "soon" and "it's on the list" vapourware speak only bears 
on your reputation.  I know.. I do the same thing :)

So if you want to get things done, you positively have to understand 
at any given point in time what is the most important thing to get 
done right now and if you're not doing it, you're not making progress 
at the fastest possible rate.
 -- Joel on Software


I'm guessing you're a 'switched on' kind of guy.. so..  .. Altme 
simply isn't one of your priorities.

I'm not really interested in debating with you, and other than this 
last statement on the topic I won't.  I'm only responding because 
there are a lot of new people here who might not be clear on the 
facts.


Vapourware is "New software that has been announced or marketed but 
has not been produced." We have been produced, people use our software, 
and it sells.  


Nor do we market or claim any of the new features people ask about. 
 People ask, and we tell them whether it is on the list, or if we 
are going to add it to the list.  We don't commit ourselves to a 
date.


You might not like these facts, but your accusations are inflammatory 
and baseless.  

The facts:


AltME came out as Beta, and we systematically brought it past version 
1.  We met our needs and goal.  If you would like something updated, 
I will give you the bank account number to transfer funds to, we 
can set a rate, and you can have any update you like.


We have never said no to any idea for AltME, we discuss them, and 
log people's ideas so that when we make changes, the ones most asked 
for are reflected (this group here on Rebol3 is not our only source 
of input).


If you do not like the rate with which we make changes, or the reason 
we make changes, then this is going to be a painful environment for 
you to remain in.


Otherwise I will state again, we are active on the development of 
an update to AltME, and we have no set date.   


If you would like to ask a constructive question I will always do 
my best to give a clear answer.  If you simply want to needle and 
dig at us, then I can no longer engage you.

You guys have done a great job with AltME. And it's free for us here! 
I can't think of any reason to complain. Quite the contrary: Thank 
you!
(If I needed something more, I should do it myself.)

despite its quirks, AltME works wonderfully and is a relaxing, non-intrusive 
way of instant messaging. being restricted to mailing lists or webforums 
would have made communications a whole lot slower.

I also received email from one earlier reboller, that rebol's image 
is hurt also because of how do we aproach development (and the person 
in question meant product which has impact - Rebol itself, altme) 
... ppl could not be simply sure what exactly is going to happen 
and when.....

we are running in circles of small adoption, open-source etc. topics, 
and how to improve current situation :-)

And I also don't use the desktop.

I will comment (for what it is worth) a little on AltMe's improvement.


If you guys have paid attention to Carl's reply on the DevCon Q&A...one 
of the things he highlighted is that /Platform is sold to SafeWorlds...(which 
is AltMe).


Draw your own conclusions there. Again, Reichart's commetns that 
they do things as they seems fit and of course $$ rules still applies.

IIRC my main concern with desktop was, there was no launch enabled, 
so it was only usable with tricks. And encapping made no sense because 
launching scripts was is prohibited. And then, after a while one 
had to search for the source, it was somewhere hidden.

About Terry, well, thats his language.. ;) That Reichart sometimes 
says "comming soon" and then it takes month i agree. OTOH altme works 
and works and recycles and works.

my ADSL is very unstable right now.... and modem is way too expensive 
for chatting/idling

Why would AltME "start" crashing.  Do you belong to more than one 
world?  Can you confirm that it is really AltME?  As in, does it 
only crash in this (big) world, and never crashes in a smaller world?

I completely understand your reasoning,

But, in our universe we:


1. Make change systematically from a giant list of features and bugs. 
 The means multiple people checking in code that is often changed 
several dozen times and overlapped by different people.


2. The once compiled, an array of computers and operating systems 
are cleaned (often reinstalled from ghosted images of hard drives) 
and sent through a complete test suite.  


3. If anything fails…the team has to be taken off what ever they 
are working on at that moment.  Then repeat the test cycle.

This is what we release as Alpha or Beta.  Gamma is a nightmare.


These realities are interleaved with all the different products we 
work on.  It means that before we do a release candidate of anything 
we have to weigh our test-team's schedule, confirm the majority of 
developers are available.  And not just for the release, but everyone 
has to hang around for the next couple of days because after release 
there is a high chance someone will find something.

Wouldn't it be just a hour of work or so
 

No…

We have a very low bug rate (we have won awards for it.)


We take it very seriously.  On one side it means things can be a 
pain in the ass (especially for us).  On the other side, what we 
do release is much better.


One thing it does cause though is for everyone to focus and gather 
up all the things they want to change at the same time.  Sort of 
like knowing that you have to put all your stuff on one boat every 
day to set sail.  If you miss the boat you have to wait a day.

Hmmm.  Interesting.  So people that get on, post, and get off produce 
a higher number?

Ok, I'm bored ..and admit it.  But a growing society this isn't ...

Would a web front-end (with similar functionality to AltME) increase 
accessibility? Or can the functionality of AltME be subsumed into 
QTask? (which is already web-based and has messaging, calender, to 
do lists, etc)


I've seen first hand how readily people will "try out" something 
online (forum, wiki, members area, etc) but shy away when asked to 
"download this software to get access". I think part of the problem 
stems from most Windows users fear of installing any software that 
is available over the net (they still trust and prefer CDs from "reputable" 
companies).

Out of the blue: how can I browse to links published on AltMe, without 
having to type then over? A left clik on the link does not do anything, 
a write click  copy the link, but is not passable into the browser, 
and the link is not selectable...
I'm using Linux Mandriva 10.2 and Firefox.

Where I am, left button attempts to launch a browser (and fails -- 
but probably due to my messed up settings). Right button copies link 
to clipboard.

Having a little trouble getting AltME to work in Ubuntu Breezy.  
I'll need to go back and make a note of specifics, but double-clicking 
the icon does nothing (rwe all checked), and running from the terminal 
gives me an error.

I had that error to under Breezy and I believe some one else too 
(check the Linux group). It helped when I upgraded to Dapper

We are working on updates for AltME, and we really want to nail this 
dot bug.


It is hard for us to see, so anything you can tell us about it will 
help.

Your connection type and platform being key.
Mine is T1/Windows MCE


One of the key questions is, when you start up a world, AltME syncs. 
 When the bug happens (which I assume means you see no groups turn 
red), did you see it sync?

Do you see anything odd happen bellow the line that says "Rebol3 
as <name>"?

Does the progress bar (the vertical space to the left of the icons) 
fill with red?

If we can see a pattern, we can nail this.

A not about the other big big bug...the recycle bug.  when I set 
my may messages to 100, I simply never see that bug.  I left my settings 
at over 1000 recently, and I knew it because I started seeing the 
recycle bug.  If this is true for others, we will change our defaults 
to start.

I have my settins at 1000, and I see the recycle bug once in a blue 
moon.

I usually see this happening when connecting for the first time, 
and clicking on a group area while AltME is syncing.

<<when I set my may messages to 100, I simply never see that bug.>>
Good tip.  I hasn't seen the recycle bug for a while.
Then, last week, I upped my message limit.
And started getting recycle crashes.
And failed to make the connection :-(

There's something deep-down that is wider than View.

Run enough CGIs under Core and you'll get what seems to be the same 
sort of error.

Which is weird as each CGI is a separate incarnation (when not using 
fastCGI, anyway).

When I saw the dot bug then most of the time a lot of the groups 
were syncing correctly and there was one or two groups that just 
didn't sync completely like it got some bad packets or something.

On the Recycle bug,  I often leave my message limit at at least 1000 
I've turned it up to 5000 a couple times because messages I was searching 
for were more than 1000 messages in the past.   If I leave it at 
5000 and then don't log in for about a week then I almost always 
get a recycle bug by logging in and clicking on the first group to 
sync while it is syncing the rest of the groups.  Waiting for syncing 
to complete will usually avoid the Recycle bug.  The number of unread 
messages in the group your switching to while it is syncing seems 
to affect whether or not you see the bug, more unread messages seems 
to increase the chance that you will see it.

(when the text is larger than the face and has to be clipped. After 
many faces it lead to a crash. ... Who posted that... )

Ideas are: 

1) timeout could create an error, leaving some resources in wrong 
state. A lot stuff in the input-buffer could trigger an buffer-overrun.

2) processing in /awake (if you use that) could not be correctly 
registered in gc. either gc misses some reference, or finds some 
reference on the return-stack which is not actually a reference. 
because the gc expects a call from a rebol-native and is instead 
called from a normal c-one.

There shouldn't be magical numbers, if you run with different data, 
and have c-level memory-bugs. Because then the memory-layout differs, 
and then the same bug hits different data. i had those bugs, perfectly 
repeatable. Disappeared when i changed the filename a bit. I guess 
different length -> slightly different memory layout. Hmm, slightly, 
sounds like an of by one in that case, or alignment?

Guys...when I say "magical," I'm speaking of testing.  I can save 
our in-house testers a lot of time if I can simpy say "set one machine 
high, one machine low. and start reducing the problem."  giving them 
a number to start with saves a lot of time.

In fact the membership aspect of Qtask and AltME protect us, but 
we have to be careful.

In response to Graham:


Yeah, I got that, and I'm saying your full of it for calling it buggy. 
 And more so, I'm saying you are mischaracterizing the issue, as 
well as what he said.  This is a pet peeve for me with you.


For the record: Carl said "they seem to work pretty well"  and they 
do!


Calling AltME buggy is simply untrue in the context, or in reply 
to what he wrote.  


It has a few bugs, but it is far form buggy.  Tomb Raider for example 
shipped with 3,000 known bugs, 280 critical crashes.  FireFox has 
more categories for bugs that AltME has bugs.  


We do agree the key bug here isn't even AltMEs, but Rebol's.  AltME 
still crashes less than any one of the other apps I run every day. 
 Thunderbird crashes twice a day on me.  I stopped using outlook 
all together.  Word has a hissy fit about 4 times a day on me.

Mike, for the record, this is a gray area.  I am in fact saying Graham 
is full of crap, but for a subjective opinion.  To actually be an 
attack of him as a person (ad hominem) I would have to say something 
about him, as opposed to about his method.  This is a rather meta-argument 
because the very thing I'm attacking is his debating skills. 


I happen to treat software as people (my own quirk), since it embodies 
the spirit of the creators.  As a result a consider some statements 
attacks.  I'm fine with a factual statement. But if I do not call 
attention to this, it becomes lore.


Gab, there is not a perceived lack of on-going development, there 
has been in fact almost no development for long stretches of time 
on AltME.  That simple.   Does not address the issue, and a disagreement 
is not a flame war.  I have too much respect for Graham to flame 
him.

Agreed.  


But if I did not engage Graham, I would be ignoring an important 
part of our relationship, which is "aside" from the point you describe. 
 


You know guys……………its OK to disagree about stuff, have deep discussions 
about them, and even walk away disagreeing to the end.  Perhaps it 
is a cultural difference that causes people to worry about "upsetting" 
people.

of course it is ok to disagree and discuss, but if we get on the 
problems too then it's even better. :)

Hey, you can hook up synapse, and have a Qtask/Altme clone.

I used to run altme all the time at work until they clamped down 
on all firewall and proxy ports. I found it to be the ideal program 
to log progress on tasks I was doing but now it wont connect. I only 
need it to do the do name or world resolution as the server would 
be running inside the firewall too. No data traffic needs to flow 
outside the firewall. Reichart - is it possible to have a local name 
world name server our enable the name resolution to be proxied through 
http, and I could just plug in the corporate gateway/proxy?

No, not yet.  We have not locked down how we want to resolve this. 
 Piracy and all.

What aspects of piracy? I just want to be able to run a world and 
clients behind the wall.

well, I'm OK with AltME. When using AltME, IRC, ICQ and Jabber, AltME 
is usually the one that connects first in a low bandwidth situation. 
But the killer feature for me is persistent chat logs, which has 
the side effect that if I post a problem, I can get response within 
minutes or hours. It's simply convenient to have running.

Guess not. What I want to be able to do is start a world inside the 
firewall and then only have clients inside it connect. The clients 
would need to know how to resolve the world name and IP address thru 
some mechanism that doesnt require going out of the firewall to do 
so or allow the name resolution and server declaration protocols 
be proxied through http. I think that should pass through most corporate 
firewalls.

actually the  AT argument could be negative or larger than 1 (which 
indicates points beyond coords)

and should also support an integer! , which would change the measure 
of  'AT to be pixels, instead of a scaling.

this is my final function, I removed the rounding as it created a 
lot of noise in position, (not the fault of the round function but 
the act or rounding itself, would snap the position back and forth 
while dragging)


I also use the vector's length using an hypothenuse formula (similar 
to distance), cause your previous example did not properly scale 
to the vector's maximum size.


I also added support for supplying a negative value for 'AT argument 
so that it actually offsets from the end, which is very usefull.

vlength: func [v] [v: v * v square-root v/x + v/y]

between: func [
	{compute the specified point on the line}
 	start [pair!]
	end [pair!]
	at [decimal! integer!]
	/local vector points 
] [
	; solve "degenerate case" first
	if equal? start end [return start]
	vector: end - start	 
	if integer? at [
		; convert AT to decimal
		at: at / (to-integer vlength vector)
	]
	if negative? at [
		at: 1 + at
	]
	

 start + to-pair reduce [to-integer vector/x * at to-integer vector/y 
 * at] 
]

the layers will allow us to optimise what gets refreshed and allow 
us to stack some properties of all graphic elements on the same layer, 
even if the graphic elements themselves are rendered as a list (of 
overlapping things)

Hi, I have a question concerning color palette conversions.


I have a color palette of greens that I would like to convert to 
blue using the same "look and style" as the green ones. Is this possible?

Robert, convert your colors to HSV and change Hue.

It's stated that this palette is specialy designed to harmonize. 
And I want to get a palette being based on blue with the same properties.

using HSV it is trivial to divide your start and end points into 
as many steps as you need

'hsv-to-rgb and 'rgb-to-hsv are in 2.7.6

Thanks guys, I know about the hsv-rgb conversion functions, and the 
interpolation can be improved using them, but I really wanted to 
improve the colours specified in the COLORS block. That means I'm 
looking for palettes of ~6 colours which aim to define a colour range 
like that of cooling metal "in the real world". But never mind, while 
it looks really cool to use a range of 6 colours, I'm now more of 
the opinion that it confuses the user interface too much - so I've 
now using a simple 2-colour range !
Thanks Chris for a look at your code.

No, just grading hue by itself is not enough to qualify as "scientific 
data" for me :) You understand that different materials have different 
colours at different temperatures. As gold and steel cool down from 
hot white, they must pass through various colours and brightnesses 
before returning to their normal room temperature colours. This "cooling 
colour signature" varies from material to material, obviously. That's 
what I meant. But anyway, I'm likely to get further using my "artistic 
eye" than searching for such data.

What different materials differ on would be which wavelength they 
absorb, while emitting. (and clearly reflection has a part in that), 
but for metal I'd guess that most of the light you see is coming 
from the surface. I don't think these things are significant enough 
for your purposes, but maybe they are, i'd need to check. doesn't 
Wikipedia have this stuff?

I don't think I mentioned what I wanted the colours for - a progress 
bar. After consideration, I've pretty much decided that a large range 
of colours such as cooling metal might go through might actually 
be bad for the user interface (unless extreme measures are taken 
to make it really look like molten metal, like a computer game programmer 
might do), because all those colours might be confusing, and some 
of them (eg. red) may signal warning, etc. (especially if the user 
hasn't seen the progress bar before - I don't want to present the 
user with a circus of colours). Fading from white to solid blue is 
what I'm using at the moment, and I'm thinking white might be too 
bright.

May I suggest that:


- Starting with just cold metal - metal is not a single colour.  
You are seeing many colours at all times, and different patches of 
colour everywhere.  You are seeing colours reflected and you need 
a coefficient table to calculate the reflection index.  Copper for 
example would lean heavily towards red.


- Next, heated metals are the same, meaning, not a single colour 
either, but now for a different reason.  But to produce the illusion 
of something burning hotter and hotter will require something that 
is multi-pass, and changing constantly.


In video games to produce the afterburner on a jet we did a few tricks:


- The jet flame itself was made of a cone-like shape (maybe a dozen 
polygons).

- The cone was set with an alpha channel so that the source was close 
to opaque, and the tip was close to translucent.

- A second cone was placed around the first cone, but just slightly 
larger.

- They both undulated at all times.  In other words, the length was 
always changing for both cones independently just slight.  And when 
the jet went faster and slower, they changed from long to flat (with 
the plane itself).

- The textures on the polygons were already a rainbox of colours, 
but as the jet changed what it was doing, the colour pallet was changed. 
 Again, both cones were not always changed at the exact same time.

- Just behind the jet (but depending on your angle of view), several 
filters (polygons that read the bg info and render again) would be 
used to create a small waiver and a refocusing.  The more GPU you 
had, the more of this you could do, and the better the final effect 
would look.

- Lastly, and this is what makes the whole thing amazing in a 3D 
game.  We are always checking the location of bright things, such 
as the sun, or things like the after burner filling your screen. 
 If so, we change the contrast of the whole world, and flare out 
your iris.  In the case of the sun we throw up a lens flare, and 
darken all the ships in the sky.

Even as a 2D problem, you should attach this the same way.

This video I think drives this all home for you.

http://www.youtube.com/watch?v=mHL94qQgl_8&feature=related

the same probably applies wrt the way the color spreads out over 
the time it cools down and/or the distance from a single hot point.

Wow - you guys. This is very interesting, but, going too far for 
my purpose at the moment. I *really* like the idea of having the 
simplest, most robust code possible. Adding so many visual effects 
and detail is something for a computer game.

I'm mostly concerned with communicating as simply and directly as 
possible to the user the current overall progress of some operation 
(multiple file download is first application).

anyone know if/where I can find an implemented shortest path algorythm 
in rebol?  I know I can do my own... but if one exists and available, 
I'd rather use it  :-)

Downscaling in View and DRAW that way basically work the same (pure 
nearest neighbor), but it might be that View is sampling odd rows 
and DRAW is sampling even rows.

the gamma issue is a standard "senior artist" debunking test for 
high-end effects artists  hehe


tell the (often ego-filled, highly paid) artist to a simple pan a 
starlit field (a simple sky shot at night).   he will LAUGH and tell 
you that's its an easy thing to do...  and then watch him struggle 
with the results, if he doesn't know his theory.


what happens is that as a pixel is distributed across several pixels, 
it will dim. as the pixel gets closer and closer to a 1:1 position 
with another pixel it will pop right back.


you must pre-amplify the "energy" of the light before panning, and 
then bring it back down... which, depending on the image & screen 
will usually be a value between 1.7 to 2.2.

the final result, when facing a "junior" compositor is that the stars 
blink in and out of existence as the image moves... 


You get this blunder even in High-budget effect-heavy movies coming 
out of hollywood.

AGG has an issue on every aliased stroke, btw.... just look at a 
red circle on blue bg and then a blue bg on red, and you will see 
there is a 1-2 line offset in the circle.  


There is no way to fix this in R2 AGG and I think that part of the 
problem is based on similar gamma assumptions.  also RGB pixels will 
affect left/right edge too, which is probably just amplifying the 
problem.

oops  above should read: "and then a blue circle on red bg"

another thing that people do not often realize is that our eyes are 
also not linear...  you can see each 8 bit step in the dark areas 
of an image, but can skip 5 steps within the brighter areas of an 
image without noticing it.  which is why we always notice image compression 
in very dark parts of DVDs and not in the bright parts of it... 


its very surprising to me that many compression algorythms get this 
wrong straight out of the bat.

Max, I tried to compare the circles example as you described and 
I don't see any offset here. Can you clarify?
Also you can set the gamma correction for AGG antialiasing in R2.

But yes, it looks the AGG filters(which are separated from the AA 
code) have the same gamma issue as described in the article. I believe 
we can fix that (hope it won't slow down things).